XL…大数据到智能数据：未来营销人员的燃料

FROM BIG DATATO SMART DATA: Usingdata to drivepersonalized brand experiences Rob Salkowitz, MediaPlantdirector of content and strategy,principal author and investigatorJanuary 22, 2014 ABSTRACT Thiswhite paper looks at Big Data from the marketing perspective, running throughthe opportunities and risks of data-driven market targeting, analytics, quantitativeperformance metrics, visualization, and potential backlash. IT’S NOT ABOUTTHE SIZE OF THE DATAIT’S WHATYOUDO WITH IT NEARLY EVERYONE IN MARKETING TODAY SHOULDBEFAMILIAR WITH THE CONCEPT OF BIG DATA—THE VOLUMINOUS STREAM OF DIGITAL INFORMATION GENERATEDAS A BYPRODUCT OF OUR TECH-DRIVEN, MOBILE, WIRELESS,HIGH-SPEED, NETWORKED WORLD. WHAT NEW MARKETINGMODELS DOES IT CREATE, AND WHAT ARE THE CHALLENGESTHAT REMAIN? The immediate goal of marketers is to use this informationto createpersonalized brand experiences—whether throughbetter ad targeting or better service delivery—but ultimatelyit will enable companies to deliverpersonalized productsandpersonalized pricing. At its endpoint, Big Data will givemarketers the tools to establish the mode and level ofengagement necessary to attract and retain each individualcustomer at the lowest cost, and manage the ongoingrelationship at optimum levels of profitability. The promise of Big Data is tempting for marketers, who are scrambling to infuse their existing processes withdata to improve message targeting, better track performanceof marketing investments, and anticipate new opportunitiesthrough predictive analytics. As the apocryphal client onceremarked, “I know I’m wasting half my ad budget, but I don’tknow which half.” Now, perhaps, we might be able to know. Big Data will give marketers the tools to attract andretain each individual customer at the lowest cost,and manage the ongoing relationship at optimumlevels of profitability. But is that really the right question? After all, the purpose ofcollecting and analyzing all this information is not simply toknow, but to act. The Hype and the Promise In a July 2013 report, McKinsey Global Institute estimatedthat the application of Big Data practices could generateupward of $30 billion in additional revenue in the retail sectoralone through productivity gains, improved transparency, andmore sophisticated targeting of marketing, ads, and offers.McKinsey further notes that “shifts in retailers’ market shareand profit pools can total more than $600 billion as the mosteffective companies capture consumer spending that wouldhave gone to competitors.”1 disparate sources necessary to achieve this vision, and whatrole can marketing play in what is essentially an IT challenge?What can brands offer customers to make the exchangeof data less “creepy” and more convenient? What changesmay be required of marketing organizations and marketingprofessionals as data analysis becomes more central tothe discipline? Big Data andSmart Clouds Cloud computing is a driving enabler of Big Data. Manyof the processing- and storage-intensive computingtasks around Big Data analytics are being performedin cloud-hosted environments, using software as aservice (SaaS) or platform as a service (PaaS) modelsthat charge based on usage. This brings the power ofBig Data analytics within the reach of small and mid-size enterprises, and enables startup Big Data serviceproviders to scale up quickly. Moving forward, expertsexpect cloud architectures to get smarter, takingadvantage of advances in adaptive machine learningto improve speed, performance, and reliability for datamanagement tasks. In this paper, we have gathered research and insights fromindustry experts in marketing and IT, including corporatemarketing and brand executives, agency leads, entrepreneursworking on data-oriented startups, academics, and hardwareand software systems specialists. The goal is to facilitateconversation between business leaders and IT around thereal challenges and real objectives of Big Data projectsfor marketing. Those are high stakes and lots of companies are scramblingto make sure they are on the right side of that shift. Butgetting from here to there is, as always, the tricky part. Weare still in the very earliest days of the Big Data era. Criticalissues related to technology, policy, and privacy remain tobe resolved. How can organizations bring together data from CLEARING THE HURDLES systems to the cloud, the economicand technological barriers to Big Datainsights are constantly falling lower,putting even more competitivepressure on companies to use thatdata effectively. time analytics. These are typicallybuilt on a framework called Hadoop,which scales and distributes bothstorage and processing power acrossa massive network, enabling analyticsengines to work with enormous,unstructured data sets such as socialdata, unformatted text, and richmedia, as well as large volumes ofmachine-generated data from sensors,radio-frequency identification (RFID),and industrial automation systems. BEFORE MARKETERS CAN